Conversational AI for Operational Automation & RAG vs Fine-Tuning

Hi Everyone,
We’re back with the third edition of Expert Talks for the year.
Please join us on 28 March 2026, from 10:30 AM to 1:00 PM @ Holiday Inn OMR IT Expressway - Wodeyar meeting room.

Agenda :
1. Welcome & Intros (10 mins)
2. Talk 1: RAG or Fine-Tuning? Stop Guessing, Start Knowing (60 mins)
3. Tea Break (15 mins)
4. Talk 2: Conversational AI for Operational Automation (30 mins)
5. Networking

Talk 1: RAG or Fine-Tuning? Stop Guessing, Start Knowing
by Balanagammal Murugan

Most teams building AI-powered solutions often face the same question: should we use RAG or fine-tune our model? Making the wrong choice costs time, money, and user trust.
In this session, we demystify both approaches through a live, side-by-side demo using a Customer Support Bot as our anchor use case. Attendees will see the same customer query handled three ways: by a base model, a RAG-powered bot, and a fine-tuned model and understand exactly why the outputs differ. We'll also explore the theory behind LoRA and QLoRA that makes fine-tuning accessible on consumer-grade hardware, and close with a practical decision framework every attendee can apply to their own use case.

What Attendees Will Learn

• How RAG and fine-tuning differ architecturally and when to use each
• How PEFT approaches (LoRA / QLoRA) make fine-tuning practical on consumer hardware
• The real cost, latency, and quality trade-offs between both approaches
• How to combine RAG + fine-tuning for production-grade systems
• A decision framework to self-classify their own use case

Optional: Our previous talk AI agents demystified: Understanding the pillars of autonomous intelligence offers useful background.

Talk 2: Conversational AI for Operational Automation
by Vijayakumar G

Operational teams are frequently burdened by high-volume distribution lists (DLs). This talk will demonstrate how Conversational AI and Large Language Models (LLMs) are being successfully deployed to automate the process of triaging and responding to incoming emails.
We will present the architecture of an intelligent agent designed to read typical DL emails, leveraging LLMs for accurate summarization, content tagging, and priority scoring. A core feature is the agent's ability to identify crucial missing information and autonomously generate follow-up emails for data collection. This capability significantly cuts down on human response time and dramatically boosts operational efficiency.
The session will include practical case studies, details on the underlying models, and best practices for deployment.

About the Speakers

Balanagammal Murugan is a software Engineer with 8+ years of experience having built scalable and resilient applications across mobile, web, and backend platforms. She has worked across diverse domains and tech stacks, delivering solutions across various scales and complexities. Driven by a deep curiosity for how things work, she enjoys diving beneath the surface to understand systems at a fundamental level. Passionate about AI, She is now expanding her horizons into LLMs, retrieval systems, and the science behind how models learn and evolve.

Vijayakumar G is a technology leader with 14+ years of experience specializing in enterprise automation, cloud architecture, and enterprise transformation. As Lead Consultant at Equal Experts India Pvt. Ltd, he designs scalable LLM-powered systems and multi-agent architectures that drive automation, cost optimization, and operational excellence across global organizations.